Survey of Intravitreal Injection Practice Patterns Among Retina Specialists

Abstract

Purpose:

Intravitreal injection therapy (IVT) is the most performed procedure in ophthalmology. This study was conducted to determine current trends in IVT delivery.

Methods:

An online, 31-question, multiple-choice survey was sent to 1677 retina specialists. The survey consisted of 3 sections: general questions, procedure technique, and postprocedure technique.

Results:

A total of 264 (16%) retina specialists completed the survey. The use of povidone-iodine (100%) and small-gauge needles (97%) was common, whereas ocular anesthesia was split among lidocaine gel (31%), lidocaine drops (25%), subconjunctival lidocaine (28%), and lidocaine-soaked pledgets (15%). More than 85% indicated povidone-iodine contributes to post-IVT corneal toxicity, and 12% reported that a needlestick injury to physician or staff occurred during IVT.

Conclusions:

Key areas for IVT improvement include optimized ocular anesthesia, development of a guarded needle for ocular drug delivery, and formulation of a less toxic ocular antiseptic.

Introduction

Intravitreal antivascular endothelial growth factor (anti-VEGF) therapy has transformed the treatment landscape of exudative macular degeneration, diabetic macular edema, and retinal vein occlusion. Multiple studies have shown not only the prevention of vision loss but also substantial visual acuity gains by 15 letters or more in 25% to 41% of patients treated with intravitreal anti-VEGF. ^{1 -5} The success of this therapy has made intravitreal injection therapy (IVT) one of the most commonly performed procedures in ophthalmology. Based on a conservative estimate, more than 6 million intravitreal injections were given in 2016 in the United States alone, ⁶ and this number continues to increase because of the aging population, expanding indications for anti-VEGF therapy, and the obesity and diabetes epidemics.

Despite the fast growth in IVT volume, the process of delivering IVT has not dramatically changed. The lack of level 1 evidence—including the injection setting (office, treatment room, or operating room), use of topical antibiotics, use of a mask, use of sterile gloves, type of anesthetics, and use of an eyelid speculum—has hampered the adoption of a single practice pattern for IVT, and surveys of IVT practice patterns have demonstrated considerable variation both globally and regionally. ^{7 -16} Practice patterns are often determined by regional convention, medical experience, and/or financial necessity. However, in many cases, significant variability is common even among retina specialists in the same practice.

In light of the variability of IVT delivery practice patterns and the lack of data on several aspects of the IVT process to date, we performed a survey among retina specialists. The goal of this survey was to obtain data on existing practice patterns to see whether there were trends that may prove useful as the field faces increasing strains from the rising volume of IVT.

Methods

We constructed an online, 31-question, multiple choice survey (Supplemental Material) using the University of Michigan Qualtrics interface. This study was approved by the University of Michigan Institutional Review Board. The survey was sent via email to 1677 retina specialists in the American Society of Retina Specialists network on April 4, 2018. A reminder email was sent to all retina specialists 1 week later. Respondents were given the opportunity to win one of five $100 gift cards.

Questions assessed participants’ overall practice settings, injection practice trends (eg, estimated average patient wait times), injection patterns, pre-IVT care and post-IVT care, and IVT toxicity assessment. Of the 31 questions, 25 were single response and 6 allowed multiple responses, of which 4 had the option of a write-in response.

Survey responses were collected via Qualtrics until May 15, 2018, when the survey was closed. Results were tabulated and the data were analyzed using SAS software (version 9.4) and the R language (version 3.5.0). The Fisher exact test was applied to compare proportions, and all P values in the text result from that test’s application. P values from Fisher exact tests were not corrected for multiple comparisons and should be interpreted in the context of the entire survey. No multiple regression analysis was conducted.

Results

A total of 264 (16%) retina specialists completed the survey. Most respondents were from North America (99.2%), with the remainder from Asia (0.8%). Thirteen respondents (5%) failed to respond to 1 or more questions.

Intravitreal Injection Therapy Frequency and Setting

Most specialists who responded work in a solo or group private practice (71%, 188 of 264), with the remainder in an academic (20%, 53 of 264) or hybrid (9%, 23 of 264) practice setting. The most common number of injections in a typical clinic day was 10 to 19 injections (39%, 104 of 264), whereas 20% (53 of 264) and 8% (21 of 264) of individuals administer fewer than 10 injections or more than 40 injections, respectively, daily (Figure 1).

Figure 1.

Number of intravitreal injections (IVTs) performed daily.

Most specialists favor administering IVT as a part of their regular clinic day (84%, 222 of 264) rather than in injection-only clinics (3%, 8 of 264). Slightly more than 50% of retina specialists perform IVT on 25% to 50% of their patients during a standard clinic day (54%, 143 of 264). The practice setting of the specialists who responded was not significantly associated with the percentage of patients who received injections (P = .38) (Figure 2).

Figure 2.

Percentage of patients receiving intravitreal injection by practice setting.

Almost 85% of retina specialists perform their injections in an examination room (223 of 264), with the remaining 16% using a treatment room (39 of 264) or operating room (2/264). Many clinicians perform bilateral injections as part of their daily practice (69%, 181 of 264) and have an assistant or technician aid in injection preparation (73%, 194 of 264).

Survey respondents estimated that patients typically spend from 30 to 90 minutes from check-in to check-out for an injection-only visit (75%, 198 of 264). This reported visit time was not found to be significantly associated with the number of injections performed daily (P = .91).

Ocular Anesthetic

Survey respondents reported using topical anesthetic drops only (25%, 66 of 261), subconjunctival lidocaine (28%, 73/261), lidocaine-soaked pledgets (15%, 40 of 261), or topical lidocaine gel (31%, 82 of 261) before IVT (Figure 3A). The primary anesthetic chosen by specialists was associated with the practice setting, with physicians in academic, hospital, or government centers favoring subconjunctival lidocaine (34%, 18 of 53) and those from a solo or group private practice preferring topical drops (28%, 52 of 186) or lidocaine gel (33%, 62 of 186) for primary anesthesia (P = .01) (Figure 3B). The chosen anesthetic was not significantly associated with the number of injections performed daily (P = .47).

Figure 3.

(A) Ocular anesthetic chosen by survey respondents. (B) Ocular anesthetic chosen by type of practice.

There was considerable variation among the respondents as to the length of time to wait between administration of anesthesia and IVT, with most respondents favoring 2 to 5 minutes (37%, 97 of 261) or 5 to 10 minutes (39%, 103 of 261). Most respondents cited patient experience (66%, 171 of 261) to be the primary reason for selecting their preferred method of ocular anesthesia, with 87% (227 of 261) of physicians citing it as one of the primary reasons for selection. About 74% (194 of 261) of respondents have had patients request a different or supplemental form of anesthesia.

Aseptic Technique

Many retina specialists surveyed do not wear a mask or gloves to administer injections (39%, 101 of 261), with only a fraction of respondents wearing both (8%, 21 of 261) (Table 1). Avoidance of sterile glove use was associated with increasing number of injections performed daily (P = .02). However, neither mask nor nonsterile glove use was significantly associated (P = .79 and P = .06, respectively) with number of daily injections. Most clinicians reported that they do not use a sterile drape for their injections (97%, 252 of 261).

Table 1.

Retina Specialists’ Responses Regarding Their Aseptic Technique.^a

	Frequency	Percentage
Personal protective equipment
Nonsterile gloves only	48	18
Sterile gloves only	34	13
Mask only	20	8
Mask and nonsterile gloves	36	14
Mask and sterile gloves	21	8
Neither mask nor gloves	101	39
Sterile drape use
No	252	97
Yes, always	8	3
Yes, usually	1	0.4
Preinjection preparation
PI eyelid preparation	171	66
PI application to conjunctiva	261	100
Time between conjunctival PI application and IVT, s
≤10	75	29
15	38	15
20	33	13
30	54	21
>30	61	23

Abbreviations: IVT, intravitreal injection therapy; PI, povidone-iodine.

^aTotal number of respondents (all categories): 261.

All 261 (100%) respondents reported applying povidone-iodine (PI) to the conjunctiva prior to IVT (261 of 261), and nearly two-thirds of specialists also prepare the patient’s eyelids with PI in preparation for injection (66%, 171 of 261). There was considerable variation among respondents as to the appropriate length of time between final conjunctival PI instillation and IVT, with a slight majority waiting 10 seconds or less (29%, 75 of 261) or more than 30 seconds (23%, 61 of 261). This waiting time was not found to be significantly associated with the number of injections performed daily (P = .11).

Injection Technique

Less than half of respondents surveyed (43%, 111 of 261) dilate their injection-only patients’ eyes prior to preparation for IVT. Most specialists prefer to use a 30-gauge needle (69%, 180 of 261) for intravitreal anti-VEGF injections, with the remaining specialists preferring a 32-gauge needle (28%, 72 of 261). The inferotemporal pars plana region is the most chosen injection site (63%, 165 of 261). Three-quarters of respondents place an eyelid speculum prior to IVT (75%, 195 of 261), but most do not use a caliper (74%, 194 of 261) or injection guide (95%, 249 of 261) to mark or assist in their IVT technique. The use of an eyelid speculum was not associated with the practice setting of the physician polled (P = .73).

Twelve percent of the respondents reported a history of a needlestick-related injury to themselves, their assistants, or both during an IVT visit (12%, 31 of 254). Among those surveyed, there was no significant relationship between the number of daily injections and needlestick-related injury to either the physician or the injection assistant (P = .89 and P = 0.85, respectively).

Postinjection Considerations

Most specialists do not use antibiotic drops after IVT (96%, 243 of 254). They tend to administer lubricating drops after IVT (6%, 14 of 254), recommend patients use lubricating drops after their visit (37%, 94 of 254), or both (31%, 79 of 254). Most estimate that less than 25% of their patients experience at least mild discomfort during the first 24 hours post injection (64%, 160 of 251); however, most respondents (94%, 240 of 254) have received post-IVT calls from patients owing to pain and foreign body sensation. Many physicians polled identified corneal toxicity as the primary cause of noninfectious pain after IVT (76%, 192 of 254), with a minority suggesting an underlying ocular surface disease (19%, 48 of 254) or needle-track inflammation (6%, 14 of 254) to be the source of discomfort. PI was the most listed source of after IVT corneal toxicity (88%, 223 of 254).

Conclusions

The results of this study suggest there are a few aspects of the IVT delivery process with consensus, including the use of PI (100%, 261 of 261), IVT delivery in an examination or procedure room (99%, 262 of 264), and use of a 30- or 32-gauge needle (97%, 252 of 261). In addition, post-IVT antibiotics are no longer commonly used, with only 11 respondents (4%, 11 of 254) reporting such use. Use of an eyelid speculum (75%, 195 of 261) and injection without calipers (74%, 194 of 261) were common, although in both cases 1 of 4 specialists do not use these techniques. One previously unexplored area of IVT delivery is the risk of injury to health care personnel. Our survey found a surprisingly high rate of needlestick injuries for physicians and assistants (12%, 31 of 254). Taken together, these data show a consolidated pattern of treatment-delivery protocol in several areas while identifying several areas that deserve additional focus to further optimize care.

The process of administering IVT has evolved over the last 15 years. In 2004 an expert panel presented a set of guidelines for intravitreal injections, and this included the routine use of both a sterile drape and a lid speculum for all patients. ¹⁷ These guidelines were instituted when there was a comparatively low level of familiarity with the procedure. There has been considerable evolution of the knowledge of risks associated with IVT and compensatory development of process recommendations over the last decade and a half. of particular, the occurrence of endophthalmitis has been shown to be rare regardless of the clinical setting in which IVT is performed. ^{18 -22} A meta-analysis performed in 2011 reported the average rate of post-IVT endophthalmitis to be 1 per 2000 injected eyes. ²³

One of the more notable findings of this survey was the continued lack of consensus regarding anesthesia choice. Despite the dramatic increase in IVT volume over the last decade, there has not been an obvious shift toward a single form of ocular anesthesia. In fact, we found a nearly even split among lidocaine gel (31%, 82 of 261), lidocaine drops (25%, 66 of 261), and subconjunctival lidocaine (28%, 73 of 261), with a smaller minority using lidocaine-soaked pledgets (15%, 40 of 261). Interestingly, most physicians who responded to our poll stated that patient comfort is their primary aim when choosing an anesthetic, which suggests that no form of anesthesia is clearly considered to be superior to others. Further supporting this finding, most retina specialists (74%, 194 of 261) have been asked by patients to consider different forms of anesthesia. These findings support the results of other studies that found no statistical differences in the individual anesthesia or patient-rated pain scores for lidocaine drops, lidocaine pledgets, lidocaine gel, and subconjunctival lidocaine. ^24,25

We found that different anesthesia wait times were associated with the method of anesthesia chosen. Specifically, most providers using lidocaine gel or lidocaine-soaked pledgets waited 5 to 10 minutes before IVT, whereas those using proparacaine or subconjunctival lidocaine waited only 2 to 5 minutes before IVT. Although these differences are nominal on an individual basis, taken in the context of a busy injection clinic with 50 injections daily, this could amount to more than 4 hours of cumulative time savings and resource reallocation.

Although our study confirmed that PI is universally used prior to IVT, there was variability in the amount of time providers wait between the last application of PI and IVT delivery. However, the full PI exposure time was less clear because close to 70% of respondents indicated that PI was applied at the beginning of preparation, implying that well over 2 minutes of total exposure was common. Per the Betadine label (Alcon), for full bactericidal effect, PI 5% should be applied 3 times and be allowed 2 minutes of drying time before rinsing the eye. A large study evaluating post-cataract surgery endophthalmitis isolates found that PI exposure times exceeding 15 minutes were needed to kill Staphylococcus epidermidis, the most common isolate identified in post-IVT endophthalmitis. ²⁶ Short exposure times result in incomplete bacterial killing, ^26,27 which puts patients at a higher risk for endophthalmitis. One concern with the growing volume of IVT and the need to optimize delivery speed is that shorter PI exposure times may result in higher rates of infection. Our data suggest that most providers are still allowing adequate PI exposure time.

There is also broad recognition that PI contributes to post-IVT corneal toxicity (88%, 223 of 254), with most respondents (94%, 240 of 254) having received post-IVT calls from patients because of pain and foreign body sensation. Post-IVT use of artificial tears is common to mitigate symptoms of corneal toxicity (74%, 187 of 254). These data suggest that a key area for IVT delivery would be a less toxic alternative to PI.

Interestingly, 12% (31 of 254) of specialists reported a history of needlestick-related injury to themselves or a staff member. In 2014, Shah and colleagues reported an 8% incidence of retinal physician needlestick injuries related to IVT. ²⁸ Our survey shows a slightly higher rate, with 9% (23 of 254) of physicians, 2% (5 of 254) of staff, and 1% (3 of 254) of physicians and staff affected. This number is high but becomes less surprising when considering that many physicians perform more than 2000 injections per year. The sheer volume of needle exposures highlights the risk to retina physicians and their staff. Because some intravitreal drugs need to be drawn from a vial, the needlestick injury risk is multiplied by physicians placing and removing a filter needle and a small-gauge needle for each patient. In addition, many physicians are assisted by support staff in performing intravitreal injections. Such personnel also run the risk of needlestick injury, particularly during steps in which the sharp is exposed, such as aspiration of the agent from its vial. The development of a guarded needle design and increasing use of prefilled syringes may help reduce these risks.

In conclusion, the process of delivering IVT has evolved significantly over the last 15 years. Key areas for improvement include optimized ocular anesthesia, development of a guarded needle for ocular drug delivery, and development of a less toxic ocular antiseptic.